Merge tag 's390-6.7-2' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux

Pull more s390 updates from Vasily Gorbik:

 - Get rid of s390 specific use of two PTEs per 4KB page with complex
   half-used pages tracking. Using full 4KB pages for 2KB PTEs increases
   the memory footprint of page tables but drastically simplify mm code,
   removing a common blocker for common code changes and adaptations

 - Simplify and rework "cmma no-dat" handling. This is a follow up for
   recent fixes which prevent potential incorrect guest TLB flushes

 - Add perf user stack unwinding as well as USER_STACKTRACE support for
   user space built with -mbackchain compile option

 - Add few missing conversion from tlb_remove_table to tlb_remove_ptdesc

 - Fix crypto cards vanishing in a secure execution environment due to
   asynchronous errors

 - Avoid reporting crypto cards or queues in check-stop state as online

 - Fix null-ptr deference in AP bus code triggered by early config
   change via SCLP

 - Couple of stability improvements in AP queue interrupt handling

* tag 's390-6.7-2' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux:
  s390/mm: make pte_free_tlb() similar to pXd_free_tlb()
  s390/mm: use compound page order to distinguish page tables
  s390/mm: use full 4KB page for 2KB PTE
  s390/cmma: rework no-dat handling
  s390/cmma: move arch_set_page_dat() to header file
  s390/cmma: move set_page_stable() and friends to header file
  s390/cmma: move parsing of cmma kernel parameter to early boot code
  s390/cmma: cleanup inline assemblies
  s390/ap: fix vanishing crypto cards in SE environment
  s390/zcrypt: don't report online if card or queue is in check-stop state
  s390: add USER_STACKTRACE support
  s390/perf: implement perf_callchain_user()
  s390/ap: fix AP bus crash on early config change callback invocation
  s390/ap: re-enable interrupt for AP queues
  s390/ap: rework to use irq info from ap queue status
  s390/mm: add missing conversion to use ptdescs

Author: torvalds

arch/s390/Kconfig

@@ -236,6 +236,7 @@ config S390
 	select THREAD_INFO_IN_TASK
 	select TRACE_IRQFLAGS_SUPPORT
 	select TTY
+	select USER_STACKTRACE_SUPPORT
 	select VIRT_CPU_ACCOUNTING
 	select ZONE_DMA
 	# Note: keep the above list sorted alphabetically

arch/s390/boot/ipl_parm.c

@@ -3,6 +3,7 @@
 #include <linux/init.h>
 #include <linux/ctype.h>
 #include <linux/pgtable.h>
+#include <asm/page-states.h>
 #include <asm/ebcdic.h>
 #include <asm/sclp.h>
 #include <asm/sections.h>
@@ -24,6 +25,7 @@ unsigned int __bootdata_preserved(zlib_dfltcc_support) = ZLIB_DFLTCC_FULL;
 struct ipl_parameter_block __bootdata_preserved(ipl_block);
 int __bootdata_preserved(ipl_block_valid);
 int __bootdata_preserved(__kaslr_enabled);
+int __bootdata_preserved(cmma_flag) = 1;
 
 unsigned long vmalloc_size = VMALLOC_DEFAULT_SIZE;
 unsigned long memory_limit;
@@ -295,6 +297,12 @@ void parse_boot_command_line(void)
 		if (!strcmp(param, "nokaslr"))
 			__kaslr_enabled = 0;
 
+		if (!strcmp(param, "cmma")) {
+			rc = kstrtobool(val, &enabled);
+			if (!rc && !enabled)
+				cmma_flag = 0;
+		}
+
 #if IS_ENABLED(CONFIG_KVM)
 		if (!strcmp(param, "prot_virt")) {
 			rc = kstrtobool(val, &enabled);

arch/s390/boot/startup.c

@@ -1,6 +1,7 @@
 // SPDX-License-Identifier: GPL-2.0
 #include <linux/string.h>
 #include <linux/elf.h>
+#include <asm/page-states.h>
 #include <asm/boot_data.h>
 #include <asm/sections.h>
 #include <asm/maccess.h>
@@ -57,6 +58,48 @@ static void detect_facilities(void)
 		machine.has_nx = 1;
 }
 
+static int cmma_test_essa(void)
+{
+	unsigned long reg1, reg2, tmp = 0;
+	int rc = 1;
+	psw_t old;
+
+	/* Test ESSA_GET_STATE */
+	asm volatile(
+		"	mvc	0(16,%[psw_old]),0(%[psw_pgm])\n"
+		"	epsw	%[reg1],%[reg2]\n"
+		"	st	%[reg1],0(%[psw_pgm])\n"
+		"	st	%[reg2],4(%[psw_pgm])\n"
+		"	larl	%[reg1],1f\n"
+		"	stg	%[reg1],8(%[psw_pgm])\n"
+		"	.insn	rrf,0xb9ab0000,%[tmp],%[tmp],%[cmd],0\n"
+		"	la	%[rc],0\n"
+		"1:	mvc	0(16,%[psw_pgm]),0(%[psw_old])\n"
+		: [reg1] "=&d" (reg1),
+		  [reg2] "=&a" (reg2),
+		  [rc] "+&d" (rc),
+		  [tmp] "=&d" (tmp),
+		  "+Q" (S390_lowcore.program_new_psw),
+		  "=Q" (old)
+		: [psw_old] "a" (&old),
+		  [psw_pgm] "a" (&S390_lowcore.program_new_psw),
+		  [cmd] "i" (ESSA_GET_STATE)
+		: "cc", "memory");
+	return rc;
+}
+
+static void cmma_init(void)
+{
+	if (!cmma_flag)
+		return;
+	if (cmma_test_essa()) {
+		cmma_flag = 0;
+		return;
+	}
+	if (test_facility(147))
+		cmma_flag = 2;
+}
+
 static void setup_lpp(void)
 {
 	S390_lowcore.current_pid = 0;
@@ -306,6 +349,7 @@ void startup_kernel(void)
 	setup_boot_command_line();
 	parse_boot_command_line();
 	detect_facilities();
+	cmma_init();
 	sanitize_prot_virt_host();
 	max_physmem_end = detect_max_physmem_end();
 	setup_ident_map_size(max_physmem_end);

arch/s390/boot/vmem.c

@@ -2,6 +2,7 @@
 #include <linux/sched/task.h>
 #include <linux/pgtable.h>
 #include <linux/kasan.h>
+#include <asm/page-states.h>
 #include <asm/pgalloc.h>
 #include <asm/facility.h>
 #include <asm/sections.h>
@@ -70,6 +71,10 @@ static void kasan_populate_shadow(void)
 	crst_table_init((unsigned long *)kasan_early_shadow_pud, pud_val(pud_z));
 	crst_table_init((unsigned long *)kasan_early_shadow_pmd, pmd_val(pmd_z));
 	memset64((u64 *)kasan_early_shadow_pte, pte_val(pte_z), PTRS_PER_PTE);
+	__arch_set_page_dat(kasan_early_shadow_p4d, 1UL << CRST_ALLOC_ORDER);
+	__arch_set_page_dat(kasan_early_shadow_pud, 1UL << CRST_ALLOC_ORDER);
+	__arch_set_page_dat(kasan_early_shadow_pmd, 1UL << CRST_ALLOC_ORDER);
+	__arch_set_page_dat(kasan_early_shadow_pte, 1);
 
 	/*
 	 * Current memory layout:
@@ -223,6 +228,7 @@ static void *boot_crst_alloc(unsigned long val)
 
 	table = (unsigned long *)physmem_alloc_top_down(RR_VMEM, size, size);
 	crst_table_init(table, val);
+	__arch_set_page_dat(table, 1UL << CRST_ALLOC_ORDER);
 	return table;
 }
 
@@ -238,6 +244,7 @@ static pte_t *boot_pte_alloc(void)
 	if (!pte_leftover) {
 		pte_leftover = (void *)physmem_alloc_top_down(RR_VMEM, PAGE_SIZE, PAGE_SIZE);
 		pte = pte_leftover + _PAGE_TABLE_SIZE;
+		__arch_set_page_dat(pte, 1);
 	} else {
 		pte = pte_leftover;
 		pte_leftover = NULL;
@@ -418,6 +425,14 @@ void setup_vmem(unsigned long asce_limit)
 	unsigned long asce_bits;
 	int i;
 
+	/*
+	 * Mark whole memory as no-dat. This must be done before any
+	 * page tables are allocated, or kernel image builtin pages
+	 * are marked as dat tables.
+	 */
+	for_each_physmem_online_range(i, &start, &end)
+		__arch_set_page_nodat((void *)start, (end - start) >> PAGE_SHIFT);
+
 	if (asce_limit == _REGION1_SIZE) {
 		asce_type = _REGION2_ENTRY_EMPTY;
 		asce_bits = _ASCE_TYPE_REGION2 | _ASCE_TABLE_LENGTH;
@@ -429,6 +444,8 @@ void setup_vmem(unsigned long asce_limit)
 
 	crst_table_init((unsigned long *)swapper_pg_dir, asce_type);
 	crst_table_init((unsigned long *)invalid_pg_dir, _REGION3_ENTRY_EMPTY);
+	__arch_set_page_dat((void *)swapper_pg_dir, 1UL << CRST_ALLOC_ORDER);
+	__arch_set_page_dat((void *)invalid_pg_dir, 1UL << CRST_ALLOC_ORDER);
 
 	/*
 	 * To allow prefixing the lowcore must be mapped with 4KB pages.

arch/s390/include/asm/mmu.h

@@ -11,7 +11,6 @@ typedef struct {
 	cpumask_t cpu_attach_mask;
 	atomic_t flush_count;
 	unsigned int flush_mm;
-	struct list_head pgtable_list;
 	struct list_head gmap_list;
 	unsigned long gmap_asce;
 	unsigned long asce;
@@ -39,7 +38,6 @@ typedef struct {
 
 #define INIT_MM_CONTEXT(name)						   \
 	.context.lock =	__SPIN_LOCK_UNLOCKED(name.context.lock),	   \
-	.context.pgtable_list = LIST_HEAD_INIT(name.context.pgtable_list), \
 	.context.gmap_list = LIST_HEAD_INIT(name.context.gmap_list),
 
 #endif

arch/s390/include/asm/mmu_context.h

@@ -22,7 +22,6 @@ static inline int init_new_context(struct task_struct *tsk,
 	unsigned long asce_type, init_entry;
 
 	spin_lock_init(&mm->context.lock);
-	INIT_LIST_HEAD(&mm->context.pgtable_list);
 	INIT_LIST_HEAD(&mm->context.gmap_list);
 	cpumask_clear(&mm->context.cpu_attach_mask);
 	atomic_set(&mm->context.flush_count, 0);

arch/s390/include/asm/page-states.h

@@ -7,6 +7,9 @@
 #ifndef PAGE_STATES_H
 #define PAGE_STATES_H
 
+#include <asm/sections.h>
+#include <asm/page.h>
+
 #define ESSA_GET_STATE			0
 #define ESSA_SET_STABLE			1
 #define ESSA_SET_UNUSED			2
@@ -18,4 +21,60 @@
 
 #define ESSA_MAX	ESSA_SET_STABLE_NODAT
 
+extern int __bootdata_preserved(cmma_flag);
+
+static __always_inline unsigned long essa(unsigned long paddr, unsigned char cmd)
+{
+	unsigned long rc;
+
+	asm volatile(
+		"	.insn	rrf,0xb9ab0000,%[rc],%[paddr],%[cmd],0"
+		: [rc] "=d" (rc)
+		: [paddr] "d" (paddr),
+		  [cmd] "i" (cmd));
+	return rc;
+}
+
+static __always_inline void __set_page_state(void *addr, unsigned long num_pages, unsigned char cmd)
+{
+	unsigned long paddr = __pa(addr) & PAGE_MASK;
+
+	while (num_pages--) {
+		essa(paddr, cmd);
+		paddr += PAGE_SIZE;
+	}
+}
+
+static inline void __set_page_unused(void *addr, unsigned long num_pages)
+{
+	__set_page_state(addr, num_pages, ESSA_SET_UNUSED);
+}
+
+static inline void __set_page_stable_dat(void *addr, unsigned long num_pages)
+{
+	__set_page_state(addr, num_pages, ESSA_SET_STABLE);
+}
+
+static inline void __set_page_stable_nodat(void *addr, unsigned long num_pages)
+{
+	__set_page_state(addr, num_pages, ESSA_SET_STABLE_NODAT);
+}
+
+static inline void __arch_set_page_nodat(void *addr, unsigned long num_pages)
+{
+	if (!cmma_flag)
+		return;
+	if (cmma_flag < 2)
+		__set_page_stable_dat(addr, num_pages);
+	else
+		__set_page_stable_nodat(addr, num_pages);
+}
+
+static inline void __arch_set_page_dat(void *addr, unsigned long num_pages)
+{
+	if (!cmma_flag)
+		return;
+	__set_page_stable_dat(addr, num_pages);
+}
+
 #endif

arch/s390/include/asm/page.h

@@ -164,7 +164,6 @@ static inline int page_reset_referenced(unsigned long addr)
 struct page;
 void arch_free_page(struct page *page, int order);
 void arch_alloc_page(struct page *page, int order);
-void arch_set_page_dat(struct page *page, int order);
 
 static inline int devmem_is_allowed(unsigned long pfn)
 {

arch/s390/include/asm/pgalloc.h

@@ -25,7 +25,6 @@ void crst_table_free(struct mm_struct *, unsigned long *);
 unsigned long *page_table_alloc(struct mm_struct *);
 struct page *page_table_alloc_pgste(struct mm_struct *mm);
 void page_table_free(struct mm_struct *, unsigned long *);
-void page_table_free_rcu(struct mmu_gather *, unsigned long *, unsigned long);
 void page_table_free_pgste(struct page *page);
 extern int page_table_allocate_pgste;
 

arch/s390/include/asm/setup.h

@@ -125,9 +125,6 @@ static inline void vmcp_cma_reserve(void) { }
 
 void report_user_fault(struct pt_regs *regs, long signr, int is_mm_fault);
 
-void cmma_init(void);
-void cmma_init_nodat(void);
-
 extern void (*_machine_restart)(char *command);
 extern void (*_machine_halt)(void);
 extern void (*_machine_power_off)(void);